MetaClean labels flipped

[LeonSaal]

Hi,
I am getting familiar with using patRoon 2.1.0 inside a Docker container.
I was trying to use the predictCheckFeaturesSession with the MetaCleanData::example_model and when I inspected the prediction results with checkFeatures, I noticed that mostly bad features were kept and good features removed fully:

Example: Kept features vs. removed features:
vs.

In the code the names subset with c("GOOD", "Pass") are removed:

patRoon/R/check_features.R

Lines 469 to 470 in d2923b0

	rmf <- gNames[preds[preds$Pred_Class %in% c("GOOD", "Pass"), "EIC"]]
	saveCheckSession(list(removeFully = rmf, removePartially = list()), session, fGroups[, rmf], "featureGroups")

I think it should be the other way around e.g.:

rmf <- gNames[preds[preds$Pred_Class %in% c("BAD", "Fail"), "EIC"]] 
saveCheckSession(list(removeFully = rmf, removePartially = list()), session, fGroups[, rmf], "featureGroups")

Or have I missed an important part?

Kind regards and thank you for your help!

Leon

[rickhelmus]

Hi Leon,

Many thanks for spotting this (silly!) mistake. I think it happened after refactoring some code a while ago. Anyway, it should pick the bad/fail peaks now :-)

Thanks,
Rick